[0001] The present invention relates to digital data radio receivers, and more particularly
to a digital data radio receiver with diverse antennas and method in which a data
signal preamble is rapidly evaluated to determine which of the antennas is best used
to receive the subsequent data signal.
[0002] A network of digital data radios may operate as a loosely coupled network of nodes
in which each receiver is required to rapidly acquire data transmissions without prior
knowledge of the time of transmission, frequency offset of the transmission (from
a known frequency), or the transmission mode.
[0003] A typical data transmission includes a data signal preceded by a preamble that typically
consists of a scrambled all-ones pattern, followed by a unique word. The radio receiver
is expected to acquire the preamble before attempting to demodulate the data signal
that follows. The probability of detecting and acquiring the preamble is desirably
high and the false alarm rate (declaring acquisition on noise) is desirably low, even
in a noisy environment. Further, the receiver desirably is sufficiently robust to
accommodate carrier frequency offsets that can create phase rotations of 45° per symbol
(the term symbol is used herein to refer to the units used in the preamble format,
e,
g., spread spectrum BPSK), and to accommodate rapid changes in noise level.
[0004] Receivers in such systems may have a single antenna, or several diverse antennas
to improve reception. Receivers connected to multiple antennas are expected to identify
the one of the multiple antennas that is to receive the data signal that follows the
preamble, and because the preamble has a set length the receiver must attempt to acquire
the preamble on all of the antennas and identity the best antenna within the limited
time period of the preamble.
[0005] Prior art receivers are typically too slow to operate effectively with multiple antennas,
and even with only one antenna do not operate with high probability of detection and
low false alarm rate in a noisy environment. For example, typical receivers use a
symbol length matched filter with the output acquired by a phase locked loop to remove
the offset frequency of the carrier. Acquisition is declared based on the amplitude
of the correlation peaks from the matched filter. The disadvantage of this approach
is that the phase locked loop is slow and may have large amounts of the jitter if
the signal level is near the noise level. To avoid this problem in multiple antenna
systems, it is known to provide parallel processing paths for the antennas, with the
attendant added cost and complexity of separate phase locked loop circuits for each
processing path.
[0006] An object of the present invention is to provide a digital data radio receiver and
method that is suitable for operation with one or multiple antennas and is able maintain
an acceptable probability of detection and false alarm rate, and to provide a digital
data radio receiver and method in which the variance of frequency offsets for symbols
in a block of symbols in the preamble and the average magnitude of the symbols in
the block of symbols are determined, and in which these values are used together to
evaluate the presence and quality of a data transmission.
[0007] Another object is to provide a digital data radio receiver and method in which the
variance of frequency offsets and average magnitude of symbols in several blocks of
symbols in the preamble are evaluated to improve the accuracy of the evaluation of
whether a detected signal is noise or a data signal preamble.
[0008] The present invention includes a method of evaluating a data signal preamble received
on an antenna to determine whether the antenna should receive the data signal following
the preamble, the preamble including a plurality of symbols, the method comprising
the steps of:
(a) determining frequency offsets from a desired frequency for each symbol in a block
of the symbols in a data signal preamble received on an antenna;
(b) determining the variance of the determined frequency offsets for plural symbols
in the block of symbols;
(c) determining the average magnitude of the symbols in the block of symbols, the
determination of average magnitude being performed in parallel with the determination
of variance of the frequency offsets; and
(d) evaluating the determined variance and the determined average magnitude for the
block of symbols to determine whether the received preamble is noise and to assess
reception quality at the antenna.
[0009] The invention also includes a radio receiver for evaluating a data signal preamble
received on two antennas so that one of the two antennas can be selected to receive
the subsequent data signal, the preamble including a plurality of symbols, the receiver
comprising:
means for determining frequency offsets from a desired frequency for each symbol in
a block of the symbols in a data signal preamble received on one of two antennas;
means for determining the variance of the determined frequency offsets for plural
symbols in the block of symbols;
means for determining the average magnitude of the symbols in the block of symbols;
means for evaluating the determined variance and the determined average magnitude
for the block of symbols to determine whether the received preamble is noise and to
assess reception quality at the one antenna;
means for comparing the evaluation results for the two antennas to determine which
one of the two antennas is to receive the subsequent data signal.
[0010] The invention will now be described, by way of example, with reference to the accompanying
drawings in which:
[0011] Figure 1 is a schematic depiction of a preamble for a data signal as broken into
blocks by the receiver.
[0012] Figure 2 is a block diagram of an embodiment of the demodulator in a digital data
radio receiver of the present invention.
[0013] Figure 3 is a partial block diagram of an embodiment of a correlator in the demodulator
of Figure 1.
[0014] An embodiment of the receiver and method of the present invention evaluates a data
signal preamble received on an antenna to determine whether the antenna should receive
the data signal following the preamble. The receiver determines frequency offsets
from a desired frequency for symbols in the preamble, determines the variance of the
frequency offsets, determines the average magnitude of the symbols, where the determination
of average magnitude may be performed in parallel with the determination of variance
of the frequency offsets, and evaluates the variance and the average magnitude to
determine whether the preamble is actually noise and to assess reception quality at
the antenna. The receiver may divide the preamble into blocks of symbols, and the
steps may be repeated on a block-by-block basis until acquisition is declared. The
evaluation results from several blocks may be used to improve the accuracy of the
determination whether the received preamble is noise.
[0015] The preamble includes a stream of symbols that the receiver may group into one or
more blocks. There are seven blocks with sixteen symbols per block in this example,
with two symbols being ignored due to ramping of the transmitter power amplifier,
and twelve symbols between the blocks and the unique word that are used for other
purposes.
[0016] In an embodiment of the invention in a two antenna diversity receiver system, the
steps may be performed first on a block of symbols received at one antenna and then
performed on the next block of symbols received at another antenna. This provides
two evaluations that may be compared to select the antenna with the best reception.
The results may be improved if the comparison is based on two consecutive blocks of
data from each antenna (consecutive blocks from one antenna being separated by a block
from the other antenna in this embodiment).
[0017] With reference to Figure 2, symbols in a received signal may be provided to a demodulator
10 for processing on a block-by-block basis (the signal is presumed to be a preamble
until proven otherwise, and is referred to herein as a preamble even though it is
more accurately a potential preamble). I and Q components are provided through analog
to digital converters 12 and 14 to correlators 16 and 18 to remove the symbol spreading
sequence (discussed further below). Outputs from correlators 16 and 18 are combined
and converted from cartesian to polar coordinate form in converter 20. The phase stream
from converter 20 is decimated to the symbol rate using timing derived from the magnitude
stream. The variance of the phase errors for the symbols in a block is determined
in calculating block 30 (discussed further below) and provided for use in determining
signal presence and quality (denoted Signal Quality 1 in Figure 2). The variance of
the phase errors is an indication of the variability of the phase offset, and would
be high if the signal were noise. A threshold variance value can be calculated using
standard statistical techniques for a known probability of detection and false alarm
rate. Variances over the threshold indicate that the signal is noise (at a known confidence
level) and that the signal is unacceptable for acquisition.
[0018] The magnitude of the output from converter 20 is also provided to integrator 32 for
non-coherent integration. Since offset frequency induced phase offsets of 45° per
symbol would significantly degrade the result if coherently combined, non-coherent
integration is used. This eliminates the need (and the hardware) in the prior art
to Doppler shift or to correct for oscillator offset prior to the correlator. The
integrated output is provided to timer 34 for determination of symbol timing and bit
sync amplitude. The bit sync amplitude is an average magnitude of the symbols in polar
coordinates and is a measure of signal amplitude that is provided for use in determining
signal presence and quality (denoted Signal Quality 2 in Figure 2).
[0019] The combination of frequency offset variance (Signal Quality 1 ) and average magnitude
(Signal Quality 2) may be used to determine whether the received signal is a preamble
for a data signal and the quality of the received signal on a particular antenna.
The frequency offset variance is a good measure of variability that is insensitive
to signal amplitude, while the average amplitude is a good measure of signal level
that is insensitive to variability. These two measures are thus independent and may
be used with a high degree of assurance that the determination of signal presence
and quality will be accurate.
[0020] Figure 3 shows correlator 16/18 that includes an adder tree of a length appropriate
for the number of symbols in a block in a data signal preamble. In this embodiment
the direct sequence acquisition signal has a short eleven chip fixed PN spreading
pattern on each DBPSK symbol, thereby permitting the use of a short correlator or
time invariant matched filter to capture the data symbols in the data signal. The
correlator provides an output that is a compressed pulse whose amplitude indicates
how strongly the input signal matches the desired signal (that is, the degree of correlation
among symbols in a block.) The input to the correlator may be sampled at twice a predetermined
chip rate to reduce losses due to analog-to-digital sampling on the chip edges. The
correlator stores each sample, but only every other input sample is used for each
correlation output sample thereby saving one-half of the correlator adder tree, as
illustrated in Figure 3. Thus, no input samples are dropped and the correlator is
implemented using half the number of multipliers that would otherwise be required.
The correlator output is at the input rate of twice per chip so that performance loss
is minimal while hardware savings are large. Figure 3 illustrates a portion of a correlator
that may be expanded as needed by adding AND gates in the serial shift registers.
The power consumed is minimized because unused adders have constant inputs. The circuit
illustrates the processing of an eleven bit PN code with twenty-two samples taken
at analog-to-digital converter 12/14 into correlator 16/18, two samples per PN code
bit or chip.
[0021] The determination of the variance of the frequency offsets uses the phase roll from
symbol to symbol which gives the frequency offset. The preamble is presumed to be
BPSK modulated by a scrambler so that data must be stripped before phase roll measurements
are taken. As discussed above, correlator output is converted to the polar coordinates.
Stripping off BPSK data in polar form may be accomplished by removing the most significant
bit (MSB) of the two's complement angle data. The remaining angles are differenced
to derive the frequency term. This term is in the form of the phase increment per
symbol, so it is independent of the number of chips per symbol. Recall that the measurements
are made on the symbols in a block of sixteen symbols in our example. After losing
one symbol in the correlator, the remaining fifteen are differenced, giving fourteen
differences. To get the phase increment per symbol, a multiplication by 16/14 (or
1 + 1/8) is performed so that a divide-by-16 (easy to implement in hardware) can be
used. This is done with shift and add stages and is easier than dividing by 14.
[0022] The signal quality evaluation may be based on the variance of the phases discussed
above. One method of determining the variance is to subtract the frequency estimate
from each estimate, square the result, and average the squares. However, this would
require two passes through the data, and the preferred method is use the standard
statistical formula for the variance:

where N = 14 in this example, and X
i is the phase offset for each symbol. Since this would require a divide-by-1 4, the
formula may be scaled to:

The 7/8 factor can implemented with 1 minus 1/8 logic that can be performed with
shifts and ADDs.
[0023] The phase roll described above may be used to seed phase locked loop 32 that will
aid in the demodulation of the subsequently received data signal. Once the preamble
has been accepted and the best antenna chosen, acquisition is declared. The measurements
already taken and stored in processor 28 may be used as initial conditions for phase
locked loop 32. The frequency offset is estimated by averaging the phase rotation
from symbol to symbol during the preamble and before phase locked loop 32 is enabled.
Frequency offset is provided to phase locked loop 32 as an initial condition to accelerate
phase lock, thereby saving time over the classical phase locked loop used in the prior
art. Additionally, a one symbol absolute phase sample is taken by processor 28 and
used to set the starting phase of a numerically controlled oscillator (NCO) 24. Once
seeded, NCO 24 may be enabled and connected to lead/lag filter 26.
[0024] The stream from converter 20 may be decimated to the symbol rate and the phase corrected
for frequency offset before PSK demodulation in demodulator 22. Phase errors from
demodulator 22 may be fed to NCO 24 through a lead/lag filter 26 to achieve and maintain
phase lock.
[0025] A digital data radio receiver and method evaluates a data signal preamble received
on an antenna to determine whether the antenna should receive the data signal following
the preamble. The receiver determines frequency offsets from a desired frequency for
each symbol in a block of symbols in the preamble, determines the variance of the
frequency offsets, determines the average magnitude of the symbols in the block of
symbols, where the determination of average magnitude may be performed in parallel
with the determination of variance of the frequency offsets, and evaluates the variance
and the average magnitude for the block of symbols to determine whether the preamble
is actually noise and to assess reception quality at the antenna. In a two antenna
diversity receiver system, these steps may be performed first on a block of symbols
received at one antenna and then performed on the next block of symbols received at
another antenna. This provides two evaluations that may be compared to select the
antenna with the best reception.
1. A method of evaluating a data signal preamble received on an antenna to determine
whether the antenna should receive the data signal following the preamble, the preamble
including a plurality of symbols, the method comprising the steps of:
(a) determining frequency offsets from a desired frequency for each symbol in a block
of the symbols in a data signal preamble received on an antenna;
(b) determining the variance of the determined frequency offsets for plural symbols
in the block of symbols;
(c) determining the average magnitude of the symbols in the block of symbols, the
determination of average magnitude being performed in parallel with the determination
of variance of the frequency offsets; and
(d) evaluating the determined variance and the determined average magnitude for the
block of symbols to determine whether the received preamble is noise and to assess
reception quality at the antenna.
2. A method as claimed in Claim 1 including the step of repeating steps (a) through (d)
until acquisition is declared, and further evaluating the determined variance for
plural blocks of symbols to improve the accuracy of the determination whether the
received preamble is noise.
3. A method as claimed in Claims 1 or 2 wherein the symbols are initially provided in
I and Q components to a pair of correlators, each of the correlators for removing
a spreading sequence in the symbols, in which the correlators sample each symbol at
twice a predetermined chip rate and the output signals from the correlators are based
on alternating ones of the samples, with the step of combining the output signals
from the pair of correlators and converting the combined output signals to polar coordinates.
4. A method as claimed in Claim 3 including the step of non-coherently integrating the
polar coordinate outputs for the symbols in the block of symbols.
5. A method of evaluating a data signal preamble received on two antennas so that one
of the two antennas can be selected to receive the subsequent data signal, the preamble
including a plurality of symbols, the method comprising the steps of:
(a) determining frequency offsets from a desired frequency for each symbol in a block
of the symbols in a data signal preamble received on one of two antennas;
(b) determining the variance of the determined frequency offsets for plural symbols
in the block of symbols;
(c) determining the average magnitude of the symbols in the block of symbols;
(d) evaluating the determined variance and the determined average magnitude for the
block of symbols to determine whether the received preamble is noise and to assess
reception quality at the one antenna;
(e) repeating steps (a) through (d) for a block of symbols in the preamble received
at the other of the two antennas; and
(f) comparing the evaluation results for the two antennas to determine which one of
the two antennas is to receive the subsequent data signal.
6. A method as claimed in Claim 5 including after step (e) the steps of repeating steps
(a) through (e) until acquisition is declared, and further evaluating the determined
variance for plural blocks of symbols received at the two antennas to improve the
accuracy of the determination whether the received preamble is noise.
7. A method as claimed in Claim 5 or 6 including the steps of providing a single receiver
for carrying out steps (a) through (d), and serially providing blocks of symbols from
alternate ones of the two antennas to the receiver, in which step (c) is carried out
in parallel with steps (a) and (b), with prior to step (c) the symbols in the block
of symbols are converted to polar coordinates.
8. A method as claimed in any one of Claims 5 to 7 wherein the block of symbols is initially
provided in I and Q components to a pair of correlators, each of the correlators providing
an output signal that indicates the correlation among the symbols in the block of
symbols, in which the correlators sample each symbol at twice a predetermined chip
rate and store each of the samples, and the output signals from the correlators are
based on alternating ones of the stored samples, with the step of combining the output
signals from the pair of correlators and converting the combined output signals to
polar coordinates.
9. A method as claimed in Claim 8 including the step of non-coherently integrating the
polar coordinate outputs for the symbols in the block of symbols.
10. A radio receiver for evaluating a data signal preamble received on two antennas so
that one of the two antennas can be selected to receive the subsequent data signal,
the preamble including a plurality of symbols, the receiver comprising:
means for determining frequency offsets from a desired frequency for each symbol in
a block of the symbols in a data signal preamble received on one of two antennas;
means for determining the variance of the determined frequency offsets for plural
symbols in the block of symbols;
means for determining the average magnitude of the symbols in the block of symbols;
means for evaluating the determined variance and the determined average magnitude
for the block of symbols to determine whether the received preamble is noise and to
assess reception quality at the one antenna;
means for comparing the evaluation results for the two antennas to determine which
one of the two antennas is to receive the subsequent data signal.